I don't think it's a big loss to make it optional, i.e., extra FreeBSD
coverage possibly not showing up near the end of the month.
(I'm a bit reluctant to move the job to GHA, as it would IMO probably
disproportionally uglify the current GHA script.)
* druntime/test: Force makefile to x86 on Apple ARM cpu (dmd does not support ARM)
* ci: Update cirrus macOS image to M1-based runners
* compiler/test/Makefile: Workaround issue 23517
* Fix Issue 22149 - TypeInfo_Struct names aren't unique, leading to botched equality semantics
By storing the mangled name, making the TypeInfo_Struct names truly
unique and more compact at the same time.
Requires https://github.com/dlang/druntime/pull/3527.
* Fix Issue 22150 - TypeInfo_Class names aren't unique, leading to botched equality semantics
By fully qualifying template arguments.
* [temp] Cirrus CI: Use same-named druntime branch
* Azure Pipelines: Try to use same-named druntime/Phobos branches for PRs originating from the official dlang repo
* Cirrus CI: Try to use same-named druntime/Phobos branches for PRs originating from the official dlang repo
This will allow to silence some deprecation message triggered in the Makefiles.
Later on, when other repositories have been migrated, we can remove support for HOST_DC.
This allows to statically set which version of FreeBSD to compile for
using DFLAGS, choices are FreeBSD 10, 11 or 12. If unset, the default is
to use the same version as the host, or fallback to version 11 if
compiling on a different host, or using an old D compiler to build DMD.
To be decided is whether there should be a dynamic way of detecting the
FreeBSD OS version as well, such as by using getosreldate().
HOST_DC is more self-descriptive I think; DMD & D_VERSION are rather
misleading for D host compiler and its version.
Also get rid of the `install_d` arg and use HOST_DC directly.
Based on infrastructure in ci.sh (also used by Semaphore CI), but
breaking up the testsuite into separate steps; I find that much better.
Also testing all 3 host compilers - DMD, LDC and GDC.
Unlike Semaphore, it doesn't rebuild itself and skips the dub package
test (not sure what that is).
A 32-bit FreeBSD job might be a worthwhile addition for the future.
